The comment by Villar et al. (1) about our article (2) on diabetes-related end-stage renal disease (ESRD) highlights the data limitations of renal registries. Our article acknowledged these as well. We reported that the age-adjusted diabetes-related ESRD incidence decreased from 1996 to 2006; among those aged <45 years, diabetes-related ESRD incidence concurrently decreased. We used the U.S. Renal Data System (USRDS) to determine the number of individuals with diabetes as the primary diagnosis who initiated ESRD therapy and National Health Interview Survey (NHIS) data to estimate the population with diagnosed diabetes. In USRDS, the physician's assessment, which may not reflect the actual ESRD etiology, determines the primary cause of ESRD. As Villar et al. pointed out, diabetes is often linked to ESRD risk factors, such as hypertension, and misclassified primary diagnoses may have affected observed trends. NHIS data are self-reported and do not include diabetes type.
We suggested that the declining trends among those aged <45 years may be related to the effect of glycemic control interventions among those with type 1 diabetes. Intensive insulin therapy reduces the risk of kidney disease in individuals with type 1 diabetes (3), and the proportion of younger individuals with diabetes achieving glycemic control has increased significantly between 1999 and 2002 and between 2003 and 2006 (4). ESRD often begins 15–20 years after developing diabetes (5); thus, individuals with ESRD aged <45 years would likely have developed diabetes by age <30. Although disease onset can occur at any age, type 1 diabetes usually strikes children and young adults. In 2001, among U.S. youth with diabetes aged <20 years, type 1 was more common than type 2 diabetes in all racial and ethnic groups except American Indians (6).
Registries such as USRDS are useful for epidemiological research and surveillance. Despite data limitations, our findings indicate encouraging trends in ESRD incidence in the diabetic population. However, we agree with Villar et al. that epidemiologic studies of ESRD and diabetes would benefit from clinical data determining ESRD etiology and the type of diabetes.
Acknowledgments
No potential conflicts of interest relevant to this article were reported.
